I observe a different behaviour than explained here: I have *Hubspot source*(version 1.4.1) and Bigquery Destionation V2 (2.0.13). I am using “*Incremental Append + Deduped*” strategy, however when the object in Deals/Contacts is deleted in the source it remains in the destination unchanged. If I do a complete refresh then data is correct. What I am doing wrong and how to get the expected behaviour?

Based on the information provided in the extracts, it seems like you're not doing anything wrong. The behavior you're observing is actually a known limitation of the Incremental Append and Incremental Append + Deduped sync modes. As per the [Incremental Data Synchronization tutorial](https://airbyte.com/tutorials/incremental-data-synchronization), delete operations are not propagated correctly in these modes. This means that if a record is deleted in the source, it will not be deleted in the destination during an incremental sync. The record will only be removed from the destination if you perform a Full Refresh sync. If you need to propagate deletions, you might want to consider using the [Change Data Capture (CDC)](https://airbyte.com/blog/change-data-capture-definition-methods-and-benefits) synchronization method, if it's supported by your source and destination. Alternatively, you could implement a system of soft deletes, although this might require significant changes to your data handling processes. Another option could be to periodically run cleanup jobs on your destination data, to remove records that no longer exist in the source. However, this would require you to have a reliable way of identifying such records. Please note that these are general suggestions and their feasibility might depend on your specific use case and data infrastructure.

Can you limit the message to a single channel? You have sent the same to a thread without been the right topic, here and #C021JANJ6TY. The behavior you have is expected. Today Airbyte can’t handle deleted records from Sources API only for CDC sources.

Thank you, I will limit next time 🙂